)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"50990ff2069f96139bca3cc116596cf3790c8b5a","unresolved":false,"context_lines":[{"line_number":7,"context_line":"[WIP] Address trunk ovsdb_handler issues"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This patch is mostly just for illustration purposes to show what I mean"},{"line_number":10,"context_line":"in my comments on the ovsdb_handler patch [1]. These changes could be"},{"line_number":11,"context_line":"squashed into [1] if needed."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"This is the current ovsdb_handler patch [1] rebased onto the latest"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"1ac06dbe_2ed73dee","line":10,"updated":"2016-08-23 01:44:02.000000000","message":"Was it tricky to make this rebased on [1] so that we\u0027d see just the differences?","commit_id":"0b8d1023bcf0323dec3094cb79892527151ee269"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"70fd41c1d61ec8508ec48b883a29a16e5876c772","unresolved":false,"context_lines":[{"line_number":7,"context_line":"[WIP] Address trunk ovsdb_handler issues"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This patch is mostly just for illustration purposes to show what I mean"},{"line_number":10,"context_line":"in my comments on the ovsdb_handler patch [1]. These changes could be"},{"line_number":11,"context_line":"squashed into [1] if needed."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"This is the current ovsdb_handler patch [1] rebased onto the latest"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"1ac06dbe_a7bdddd9","line":10,"in_reply_to":"1ac06dbe_2ed73dee","updated":"2016-08-23 17:19:18.000000000","message":"Ideally [1] would\u0027ve already been rebased by now, but I didn\u0027t want to step on someone\u0027s foot if it\u0027s already being worked on. I worked on this patch until it was functional after rebasing then posted it for review. I pinged jlibosva, and he said I\u0027m good to rebase it. The next PS should only have the changes I\u0027m illustrating.","commit_id":"0b8d1023bcf0323dec3094cb79892527151ee269"}],"neutron/services/trunk/drivers/openvswitch/agent/ovsdb_handler.py":[{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"70fd41c1d61ec8508ec48b883a29a16e5876c772","unresolved":false,"context_lines":[{"line_number":66,"context_line":"                continue"},{"line_number":67,"context_line":"            bridge_name \u003d ovs_lib.BaseOVS().get_bridge_for_iface(port_name)"},{"line_number":68,"context_line":"            if is_trunk_bridge(bridge_name):"},{"line_number":69,"context_line":"                trunk_br \u003d ovs_lib.OVSBridge(bridge_name)"},{"line_number":70,"context_line":"                trunk_br_ports \u003d trunk_br.get_ports_attributes("},{"line_number":71,"context_line":"                    \u0027Interface\u0027, columns\u003d[\u0027name\u0027, \u0027external_ids\u0027],"},{"line_number":72,"context_line":"                    if_exists\u003dTrue)"},{"line_number":73,"context_line":"                for trunk_br_port in trunk_br_ports:"},{"line_number":74,"context_line":"                    # don\u0027t consider the port named like the bridge"},{"line_number":75,"context_line":"                    if trunk_br_port[\u0027name\u0027] !\u003d port_name:"},{"line_number":76,"context_line":"                        self.create_trunk(trunk_br, **trunk_br_port)"},{"line_number":77,"context_line":"        for port_event in ovsdb_events[\u0027removed\u0027]:"},{"line_number":78,"context_line":"            iface_id \u003d self.br_int.portid_from_external_ids("}],"source_content_type":"text/x-python","patch_set":1,"id":"1ac06dbe_3aa58046","line":75,"range":{"start_line":69,"start_character":16,"end_line":75,"end_character":58},"updated":"2016-08-23 17:19:18.000000000","message":"I might be able to clean this logic up, because all it\u0027s supposed to do is find the vif port, which we have as port_event here already.","commit_id":"0b8d1023bcf0323dec3094cb79892527151ee269"},{"author":{"_account_id":4187,"name":"Ryan Tidwell","email":"rtidwell@suse.com","username":"ryan-tidwell"},"change_message_id":"b8cd9cf55a795a2f27c864259dcfee750778bffb","unresolved":false,"context_lines":[{"line_number":77,"context_line":"        for port_event in ovsdb_events[\u0027removed\u0027]:"},{"line_number":78,"context_line":"            iface_id \u003d self.br_int.portid_from_external_ids("},{"line_number":79,"context_line":"                port_event[\u0027external_ids\u0027])"},{"line_number":80,"context_line":"            self.remove_trunk_port(iface_id)"},{"line_number":81,"context_line":""},{"line_number":82,"context_line":"    def remove_trunk_port(self, iface_id):"},{"line_number":83,"context_line":"        ctx \u003d context.get_admin_context()"}],"source_content_type":"text/x-python","patch_set":1,"id":"1ac06dbe_e5d0a8f8","line":80,"updated":"2016-08-22 23:43:57.000000000","message":"See L84. As written, this looks like it will introduce an RPC round-trip on every port delete. We would want to filter calls to remove_trunk_port somehow. We may still want to tag the external_ids with bridge_name or some other field to indicate that we\u0027re dealing with a trunk before we proceed.","commit_id":"0b8d1023bcf0323dec3094cb79892527151ee269"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"70fd41c1d61ec8508ec48b883a29a16e5876c772","unresolved":false,"context_lines":[{"line_number":77,"context_line":"        for port_event in ovsdb_events[\u0027removed\u0027]:"},{"line_number":78,"context_line":"            iface_id \u003d self.br_int.portid_from_external_ids("},{"line_number":79,"context_line":"                port_event[\u0027external_ids\u0027])"},{"line_number":80,"context_line":"            self.remove_trunk_port(iface_id)"},{"line_number":81,"context_line":""},{"line_number":82,"context_line":"    def remove_trunk_port(self, iface_id):"},{"line_number":83,"context_line":"        ctx \u003d context.get_admin_context()"}],"source_content_type":"text/x-python","patch_set":1,"id":"1ac06dbe_fd8762b2","line":80,"in_reply_to":"1ac06dbe_e5d0a8f8","updated":"2016-08-23 17:19:18.000000000","message":"Ah, right, this would probably throw an error if another vif port is removed. I didn\u0027t try that out, but based upon your finding it does look like we need to filter it somehow. I see a couple options right now:\n\n1) set bridge_name in the external_ids of the vif port after L68.\n2) check if the OVS bridge named \"utils.gen_trunk_br_name(iface_id)\" exists. If it does, then we do the RPC call and start tearing down the trunk.\n\nOption 2 saves us from having to set bridge_name in the external_ids, but it requires the logic of checking that a trunk bridge exists, which will be redone in the trunk_manager anyways.\n\nI think I\u0027m leaning more towards Option 1 now, but might there be another reason to keep the trunk bridge named using the port_id rather than trunk_id?","commit_id":"0b8d1023bcf0323dec3094cb79892527151ee269"}],"neutron/services/trunk/drivers/openvswitch/driver.py":[{"author":{"_account_id":4187,"name":"Ryan Tidwell","email":"rtidwell@suse.com","username":"ryan-tidwell"},"change_message_id":"b8cd9cf55a795a2f27c864259dcfee750778bffb","unresolved":false,"context_lines":[{"line_number":69,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details.\"\"\""},{"line_number":70,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":71,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":72,"context_line":"        set_br_name(utils.gen_trunk_br_name(port[\u0027id\u0027]))"}],"source_content_type":"text/x-python","patch_set":1,"id":"1ac06dbe_a50fd086","line":72,"updated":"2016-08-22 23:43:57.000000000","message":"I see you\u0027re switching from using the trunk_id to port id for generating the trunk bridge name. I\u0027m thinking that we still need to stick trunk metadata in the external_ids anyway. I\u0027m not sure this buys us a whole. This change seems more cosmetic than anything.","commit_id":"0b8d1023bcf0323dec3094cb79892527151ee269"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"70fd41c1d61ec8508ec48b883a29a16e5876c772","unresolved":false,"context_lines":[{"line_number":69,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details.\"\"\""},{"line_number":70,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":71,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":72,"context_line":"        set_br_name(utils.gen_trunk_br_name(port[\u0027id\u0027]))"}],"source_content_type":"text/x-python","patch_set":1,"id":"1ac06dbe_f898106c","line":72,"in_reply_to":"1ac06dbe_a50fd086","updated":"2016-08-23 17:19:18.000000000","message":"Yes, this is mostly cosmetic, but I\u0027m trying to think of other reasons as to why we might want the trunk bridge named using the port_id rather than the trunk_id. The patch ports are all named using the port_ids, the vif port uses the port_id, but here the trunk bridge using trunk_id is the odd man out.","commit_id":"0b8d1023bcf0323dec3094cb79892527151ee269"}]}
